Interannual variation of springtime biomass burningin Indochina: Regional differences, associatedatmospheric dynamical changes,and downwind impactsWan-Ru Huang1, Sheng-HsiangWang2, Ming-Cheng Yen2, Neng-Huei Lin2, and Parichart Promchote3,4

1Department of Earth Sciences, National Taiwan Normal University, Taipei, Taiwan, 2Department of Atmospheric Sciences,National Central University, Taoyuan, Taiwan, 3Department of Agronomy, Kasetsart University, Bangkok, Thailand,4Department of Plants, Soils, and Climate, Utah State University, Logan, Utah, USA

Abstract During March and April, widespread burning occurs across farmlands in Indochina inpreparation for planting at the monsoon onset. The resultant aerosols impact the air quality downwind. Inthis study, we investigate the climatic aspect of the interannual variation of springtime biomass burningin Indochina and its correlation with air quality at Mt. Lulin in Taiwan using long-term (2005–2015) satelliteand global reanalysis data. Based on empirical orthogonal function (EOF) analysis, we find that the biomassburning activities vary with two geographical regions: northern Indochina (the primary EOF mode) andsouthern Indochina (the secondary EOF mode). We determine that the variation of biomass burning overnorthern Indochina is significantly related with the change in aerosol concentrations at Mt. Lulin. This occursfollowing the change in the so-called India-Burma Trough in the lower and middle troposphere. When theIndia-Burma Trough is intensified, a stronger northwesterly wind (to the west of the trough) transports thedryer air from higher latitude into northern Indochina, and this promotes local biomass burning activities.The increase in upward motion to the east of the intensified India-Burma Trough lifts the aerosols, whichare transported toward Taiwan by the increased low-level westerly jet. Further diagnoses revealed theconnection between the India-Burma Trough and the South Asian jet's wave train pattern as well as theprevious winter's El Niño–Southern Oscillation phase. This information highlights the role of the India-BurmaTrough in modulating northern Indochina biomass burning and possibly predicting aerosol transport to EastAsia on the interannual time scale.

1. Introduction

During March and April, widespread agro-residue burning occurs across farmlands in Indochina (red area inFigure 1a) in preparation for planting at the monsoon onset. This burning occurs in conjunction with localhot-dry-stagnant air and generates smoke and haze aerosol pollution, particularly over the valley terrain ofthe northern peninsula [Kim Oanh and Leelasakultum, 2011]. Because the resultant aerosols affect not onlythe local region but also the air quality downwind, many studies have examined the climate and weathercharacteristics in association with biomass burning [Duncan et al., 2003; Yen et al., 2013; Wang et al., 2015].In general, biomass burning occurs over Indochina during the time period from February to April, with a max-imum occurrence in March (Figure 1b). After the monsoon rainfall onset in late April, the occurrence of fireevents becomes minimal. The annual variation of aerosol concentrations over Indochina depends on the pre-vailing times of biomass burning [Gautam et al., 2013; Huang et al., 2013]. Some reports revealed that aerosolsgenerated by biomass burning could be reduced by the increase of precipitation and monsoon circulationdue to rainout and washout processes [Sanap and Pandithurai, 2015; Sonkaew and Macatangay, 2015].

Based on the NOAA's Hybrid Single-Particle Lagrangian-Integrated Trajectory model simulations, earlier stu-dies [Yu et al., 2008; Yen et al., 2013] demonstrated that the long-range transport of Indochina's biomass burn-ing pollutants has a significant impact on the surface air quality of downstream areas, particularly in Taiwan.The biomass burning emissions are uplifted above the planetary boundary layer through regional conver-gence at the lower level and strong thermal buoyancy [e.g., Yen et al., 2013; Wang et al., 2015]. After beinglifted, the aerosols are transported to Taiwan by the low-level westerly jet (Figures 1c and 1d) within 2–3 days[e.g., Cheng et al., 2013; Yen et al., 2013; Dong and Fu, 2015]. These aerosol transports can result in significantincrease of air pollutant concentrations in Taiwan, as reported by many studies [e.g., Lee et al., 2011; Yen et al.,

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,016

PUBLICATIONSJournal of Geophysical Research: Atmospheres

RESEARCH ARTICLE10.1002/2016JD025286

Key Points:• The springtime biomass burningactivities in Indochina vary with twogeographical regions: northernIndochina and southern Indochina

• The interannual variation of biomassburning over northern Indochina isclosely related with the changes ofPM10 at Mt. Lulin in Taiwan

• The changes of the India-BurmaTrough are important to themodulation of the aerosol transportpattern from northern Indochina toTaiwan

Correspondence to:W.-R. Huang,wrhuang@ntnu.edu.tw

Citation:Huang, W.-R., S.-H. Wang, M.-C. Yen,N.-H. Lin, and P. Promchote (2016),Interannual variation of springtimebiomass burning in Indochina: Regionaldifferences, associated atmosphericdynamical changes, and downwindimpacts, J. Geophys. Res. Atmos., 121,10,016–10,028, doi:10.1002/2016JD025286.

Received 27 APR 2016Accepted 14 AUG 2016Accepted article online 17 AUG 2016Published online 5 SEP 2016

©2016. The Authors.This is an open access article under theterms of the Creative CommonsAttribution-NonCommercial-NoDerivsLicense, which permits use and distri-bution in any medium, provided theoriginal work is properly cited, the use isnon-commercial and no modificationsor adaptations are made.

2013; Chuang et al., 2015]. For example, Chuang et al. [2015] noted that the concentrations of particulatematter (PM)2.5 and PM10 at Mt. Lulin of Taiwan was increased from approximately 15–20μgm�3 to53–64μgm�3 on 12 March 2010 after the arrival of the biomass burning plume from Indochina.

Even though the aforementioned relationship between biomass burning in Indochina and its downwindinfluences on concentration of air pollutants (e.g., PM10 and PM2.5) in Taiwan has been reported [e.g., Yenet al., 2013; Chuang et al., 2015], the contribution of the low-level trough over the Bay of Bengal (so-calledIndia-Burma Trough; Figure 1d) on the aerosol transport pattern has not been emphasized. Several studieshave suggested that the India-Burma Trough is a key system affecting the weather and climate overSoutheast Asia, especially in winter [e.g., Suo et al., 2008; Suo and Ding, 2009; Zhou et al., 2009; Zong et al.,2012;Wang et al., 2011; Li and Zhou, 2016]. Suo and Ding [2009] noted that the India-Burma Trough is gener-ally established in October, is maintained from November to the following February, and peaks fromMarch toMay. Li and Zhou [2016] further suggested that the interannual variation of winter precipitation over south-west China is modulated by the change of the India-Burma Trough. As inferred from this documented litera-ture, it is likely that the changes in the India-Burma Trough might play an important role in modulating theinterannual variation of biomass burning in Indochina and its downwind impacts on the air quality in Taiwan.

In 2007, the Seven SouthEast Asian Studies (7-SEAS) project (http://7-seas.gsfc.nasa.gov/) was established toperform research in the field of aerosol-meteorology and climate interactions from Java through the MalayPeninsula and Southeast Asia to Taiwan [Reid et al., 2013; Lin et al., 2013]. The Lulin AtmosphericBackground Station (LABS), located at Mt. Lulin (altitude: 2862m; 23°28′07″N, 120°52′25″N; http://lulin.tw)in central Taiwan, started operations on April 2006 to provide continuous observations of a variety of aerosol

Figure 1. (a) Geographic location of Indochina (red marked area). The location of the Lulin Atmospheric Background Station (LABS; altitude: 2862m) in Taiwan ismarked by the pink dot in Figure 1a. (b) Area-averaged monthly mean rainfall (blue bar) and biomass burning amount (i.e., MODIS fire count; red line) overIndochina. (c) Climatological mean of March and April (hereafter, MA mean) wind vectors at 850 hPa and the MAmean of precipitation (shaded). (d) MA mean of the700 hPa wind vectors and biomass burning amount (shaded). The blue line in Figure 1d represents the India-Burma Trough.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,017

chemistry parameters, trace gases, etc. The long-term observational data collected by the 7-SEAS project andLABS in Taiwan now provide a great opportunity for studying the relationship between the variations of bio-mass burning in Indochina and the variations of air quality in Taiwan on an internal interannual time scale.

This study aims to examine the interannual variation of biomass burning in Indochina and its associatedatmospheric dynamical changes, with a focus on understanding the role of the India-Burma Trough in mod-ulating the aerosol transport from Indochina to LABS in Taiwan. Considering a large domain from the tropicsto 25°N of Indochina (Figure 1a), the atmospheric circulation features important for the modulation of theinterannual variation of biomass burning in different subregions of Indochina might be different. This infer-ence, together with the regional differences in the interannual variation of biomass burning and its down-wind impacts, will be clarified in the subsequent analysis.

The remainder of this paper is organized as follows. In section 2, the data used for the analyses are described.Section 3 reports the spatial-temporal characteristics of the interannual variation of biomass burning inIndochina, their relations to air quality in Taiwan, and the associated atmospheric circulation features.Discussion is provided in section 4. A summary is given in section 5.

2. Data and Methodology

In this study, the biomass burning activity from 2005 to 2015 is identified by fire data from ModerateResolution Imaging Spectroradiometer (MODIS; http://modis.gsfc.nasa.gov/) on board NASA's Terra andAqua satellites. The MOD14 (Terra) and MYD14 (Aqua) active fire product detects fires in 1 km pixels thatare burning at the time of overpass under relatively cloud-free conditions [Giglio et al., 2003]. The overpasstime for Terra and Aqua are approximately 10:30 and 13:30 local time, respectively. The daily fire countsare calculated from the sum of the MOD14 and MYD14 data sets. Because the MODIS fire count data exhibitsome quality problem in the early 2000s [Giglio, 2013; Giglio et al., 2013], only data from 2005 onward wereused. For a detailed review of the MODIS fire count data, please refer to Giglio [2013].

To further validate the representativeness of MODIS fire count for biomass burning activities over Indochina,we also examined the biomass burning dry matter emissions (hereafter, biomass burning emissions) pro-vided by the Global Fire Emissions Database, version 4 (GFED4s; http://globalfiredata.org/data.html) [Giglioet al., 2013]. The GFED4s provides various types of biomass burning emissions, and this study only usedthe changes of total amount of biomass burning emissions (i.e., the summation of all types). It should benoted that MODIS exhibits a limit in capturing peat burning [Giglio et al., 2010]; however, we considered thislimitation neglectable because the proportion of peat burning to total biomass burning emissions overIndochina is small [e.g., Shi and Yamaguchi, 2014]. As will be shown in Figure 3 (discussed later), the interann-ual variation of the GFED4s' total biomass burning emissions matches well with that of the MODIS fire countover Indochina. This result indicates that MODIS fire count can represent biomass burning activities overIndochina, consistent with Yu et al. [2008] and Wang et al. [2015].

To examine the impact of biomass burning on downwind air quality, we selected LABS as the representativelocation because the transportation of the biomass burning plume from Indochina to Taiwan has been welldocumented by the measurement data collected from this station [e.g., Ou Yang et al., 2014, and referencestherein]. The instrument for the PM10 and PM2.5 mass concentration is the tapered element oscillatingmicrobalance (RP 1400a, USA). In this study, we used only PM10 data (~10 years) because of its longer recordcompared to PM2.5 data (~3 years).

For the observational precipitation, we used 3 hourly Tropical Rainfall Measuring Mission (TRMM) 3B42 satel-lite precipitation [Huffman et al., 2007]. TRMM 3B42 data have been widely used for the depiction of rainfallvariation over South and East Asia [Hong et al., 2005; Zhou et al., 2008; Huang and Chan, 2012; Huang andWang, 2014]. For the meteorological data (including wind fields and humidity), we used the NationalCenters for Environmental Prediction–Department of Energy (NCEP-DOE) Reanalysis 2 [Kanamitsu et al.,2002]. The calculation of the moisture flux from NCEP-DOE Reanalysis 2 is based on equation (1):

Q ¼ 1g

∫ps

300Vq dp

� �; (1)

where V denotes the horizontal wind vectors, q is the specific humidity, g is the gravity, and p is the pressurelevel [Chen et al., 1988]. Following Chen et al. [1988], the moisture flux can be separated into rotational (QR)

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,018

and divergent (QD) components, which can be further expressed in terms of the stream function (ψQ) andpotential function (χQ) of the moisture flux:

Q ¼ QR þ QD ¼ k̂ � ∇ψQ þ ∇χQ; (2)

∇2ψQ ¼ k̂ · ∇ � Q; (3)

∇2χQ ¼ ∇ · Q: (4)

The fields of ψQ and χQ are obtained by solving the Poisson equation for equations (3) and (4).

Hereafter, unless noted otherwise, the variables used for examination are the monthly means of March andApril (MA mean) during the time period of 2005–2015. The anomaly of a variable designates the removal ofclimatological monthly mean from each month. The index of northern Indochina biomass burning activitieswas defined as the amount of MODIS fire count over the domain of 17–22°N, 97–102°E (red box in Figure 2a),while the index of southern Indochina biomass burning activities was defined as the amount of MODIS firecount over the domain of 10–15°N, 102–107°E (red box in Figure 2b). The selection of these two specificdomains was based on the analysis in Figure 2, as explained later.

3. Results3.1. Characteristics of Biomass Burning in Indochina and Its Downwind Impact

Figure 1d shows the climatological mean distribution of biomass burning activities (i.e., the MODIS fire count)in Indochina averaged over the focused time period. Visually, the amount of biomass burning activities is

Figure 2. (a) The first-principle mode of the empirical orthogonal function (EOF) analysis on the anomalies of springtime(i.e., MA mean) biomass burning activities during 2005~2015. The eigenvector and eigencoefficient are given in the leftand right plots, respectively. (b) Similar to Figure 2a but for the second-principle mode. In E1 (E2), the domain of themaximum variation center in northern (southern) Indochina covering 17–22°N, 97–102°E (10–15°N, 102–107°E) is marked.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,019

larger over northern Indochina, covering most of Burma and northern Thailand, than over southernIndochina. To clarify how these biomass burning activities varied on the interannual time scale, we appliedempirical orthogonal function (EOF) analysis on the anomalies of biomass burning activities for the time per-iod of 2005–2015 (Figure 2). The EOF analysis was based on singular-value decomposition (SVD) approach.The SVD approach, which avoids having to compute the covariance matrix directly, provides an optimalway for data sets with a large spatial dimension [Hannachi et al., 2007]. Any missing space/time data wereremoved before applying the SVD approach. For the detailed review of the EOF analysis (including historicalbackground, formulation and computation, and application), please refer to Hannachi et al. [2007].

In Figure 2, only the results of the first two EOF modes, together explaining approximately 90% of the totalinterannual variability of biomass burning activities, are presented. Spatially, the first EOF mode shows thatthe variation is mainly over northern Indochina, with a largest variation center over northern Thailand(Figure 2a). Corresponding to this spatial distribution, the temporal variation of Figure 2a shows that more(less) biomass burning likely occurred over northern Indochina in the years of 2007 and 2010 (2005, 2008,2011, and 2015). In contrast, it can be inferred from the spatial-temporal patterns of the second EOF mode(Figure 2b) that over southern Indochina more (less) biomass burning occurred in the years of 2005, 2010,and 2013–2015 (2006, 2007, 2009, and 2012). Additionally, Figure 3 shows the time series of biomass burningactivities over the major variation domain of northern Indochina (17–22°N, 97–102°E; marked in Figure 2a)and southern Indochina (10–15°N, 102–107°E; marked in Figure 2b). Consistent with that implied fromFigure 2, the results of Figure 3 confirm that the interannual variation of biomass burning activities overnorthern Indochina is different to that over southern Indochina.

To understand which subregion of biomass burning has a larger impact on the aerosol concentration inTaiwan, we further compared the time series of PM10 at LABS in Taiwan (blue lines) with the time series ofbiomass burning activities (bars)/emissions (green dashed lines) over the major variation domain of northernIndochina (Figure 3a) and southern Indochina (Figure 3b). Notably, the fluctuation of PM10 at LABS in Taiwanseems to match better with the fluctuation of biomass burning activities/emissions over northern Indochinathan over southern Indochina. More specifically, the maximum (minimum) of PM10 at LABS in Taiwanoccurred in the years of 2007 and 2010 (2008, 2011, and 2015), when the maximum (minimum) of biomassburning activities/emissions occurred over northern Indochina. Statistically, the temporal correlation coeffi-cient between the interannual variation (i.e., linear trend removed) of PM10 at LABS in Taiwan and the bio-mass burning activities/emissions over northern Indochina is approximately 0.86/0.81 (significant at the95% confidence interval). Such a temporal variation of PM10 is found to be less related to the temporal varia-tion of biomass burning activities/emissions over southern Indochina (the correlation between the two timeseries in Figure 3b is only �0.15/�0.14). These findings (Figures 2 and 3) highlight the importance of

Figure 3. Time series of MA mean of biomass burning activities (bar) and total biomass burning dry matter emissions(green dashed line) over the two selected domains: (a) northern Indochina (17–22°N, 97–102°E; marked in Figure 2a)and (b) southern Indochina (10–15°N, 102–107°E; marked in Figure 2b). In Figures 3a and 3b, the blue line represents thetime series of MA mean of PM10 at LABS in Taiwan.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,020

considering regional differences in the discussions of the remote impact of Indochina's biomass burning onthe air quality in Taiwan.

3.2. The Associated Atmospheric Circulation Features

Next, we examined the atmospheric circulation changes that are associated with the change of biomassburning over northern Indochina. Here the index of northern Indochina biomass burning activities (i.e., theamount of MODIS fire count over the domain of 17–22°N, 97–102°E) is correlated with the followingassociated fields: total biomass burning emissions (Figure 4a), 850 hPa wind circulation (Figure 4b), 700 hPawind circulation (Figure 4c), stream function of the moisture flux ψQ (Figure 4d), potential function of themoisture flux χQ (Figure 4e), and precipitation (Figure 4f). It is noted that the increase in biomass burningactivities/emissions over northern Indochina (Figure 4a) is associated with the intensification of the India-Burma Trough in the lower and middle troposphere (Figures 4b and 4c). In conjunction with such circulationchange, both the northwesterly wind behind (to the west) the India-Burma Trough and thewesterly/southwesterly wind ahead (to the east) of the India-Burma Trough are intensified relative to the

Figure 4. Temporal correlation coefficient between the MA mean index of northern Indochina biomass burning activities (i.e., the amount of MODIS fire count overthe domain of 17–22°N, 97–102°E; red boxed area) and six selected variables: (a) GFDE4s total biomass burning dry matter emissions, (b) 850 hPa wind circulation V(850 hPa), (c) 700 hPa wind circulation V (700 hPa), (d) stream function of the moisture flux ψQ, (e) potential function of the moisture flux χQ, and (f) precipitation Pextracted from TRMM observations. The time period of correlation is from 2005 to 2015. Only the areas with significant changes (at the 90% confidence interval) areshaded/shown. The color scale of Figures 4a–4f is given in the bottom plot.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,021

climatological mean in the lower and middle troposphere (Figures 4b and 4c). A further examination on theassociated change of ψQ (Figure 4d) has indicated that the increase in northwesterly wind behind the India-Burma Trough could advect dryer air from the higher latitude to northern Indochina. This would facilitate amoisture divergence center to be formed over the west coast of Indochina, as revealed by the associatedchange of χQ (Figure 4e). Furthermore, such a suppression of the moisture supply from the Bay of Bengalwould result in less precipitation to be formed over northern Indochina (Figure 4f), and this dryer conditioncan help promote the local biomass burning activities (Figure 4a). The argument that less precipitation favorsmore biomass burning is also suggested by Dong and Fu [2015] but for the change of biomass burning accu-mulated over “entire” Indochina.

Figure 5 presents the correlation maps between the index of southern Indochina biomass burning activities(i.e., the amount of MODIS fire count over the domain of 10–15°N, 102–107°E) and the six selected fields usedfor northern Indochina (Figure 4). It is noted that the increase in biomass burning activities/emissions oversouthern Indochina (Figure 5a) is associated with the appearance of an anticyclone system over the Bay of

Figure 5. Similar to Figure 4 but for the temporal correlation coefficient between the MA mean index of southern Indochina biomass burning activities (i.e., theamount of MODIS fire count over the domain of 10–15°N, 102–107°E; boxed area) and six selected variables: (a) GFDE4s total biomass burning dry matteremissions, (b) V (850 hPa), (c) V (700 hPa), (d) ψQ, (e) χQ, and (f) precipitation P extracted from TRMM observations. The time period of correlation is from 2005 to 2015.Only the areas with significant changes (at the 90% confidence interval) are shaded/shown. The color scale of Figures 5a–5f is given in the bottom plot.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,022

Bengal in the lower and middle troposphere (Figures 5b and 5c). In conjunction with the Bay of Bengal antic-yclone system, the northerly wind is intensified over southern Indochina. Such a circulation change couldadvect the dryer air from higher latitude to southern Indochina, as noted from the associated change of ψQ

(Figure 5d). As a result, the moisture flux is divergent over south of the South China Sea (Figure 5e), and lessprecipitation has occurred over southern Indochina (Figure 5f). Obviously, these circulation changes that areimportant for the modulation of biomass burning over southern Indochina (Figure 5) are very different tothose important for the modulation of biomass burning over northern Indochina (Figure 4). This findingexplains why the interannual variation of biomass burning over Indochina is locally dependent.

It should be mentioned that to be transported to Taiwan, the air pollution created by Indochina's biomassburning has to be uplifted to the level with a stronger westerly jet [e.g., Yen et al., 2013; Dong and Fu,2015]. Thus, to clarify why the biomass burning over northern Indochina (as compared to that over southernIndochina) has a larger impact on the PM10 in Taiwan, we further examined the relationship between thechange in vertical motion (�ω; positive value denotes upward motion) and the change in index of northernIndochina biomass burning activities (defined in section 2) The result (Figure 6a) shows that the change innorthern Indochina's biomass burning activities is positively correlated to the change in upward motion overthe area between Indochina and Taiwan at levels up to 700 hPa. This finding, together with Figures 4e and 4f,implies that more active but dryer upwardmotion occurred over northern Indochina in the years with greaterbiomass burning. After the air pollution is uplifted to upper levels, the increase in the westerly wind ahead ofthe intensified India-Burma Trough (e.g., Figures 4b and 4c), which appeared in the levels up to 500 hPa

Figure 6. Temporal correlation coefficient between the MA mean index of northern Indochina biomass burning activities(defined in Figure 4) and two selected variables: (a) multilevel vertical motion (�ω; the positive values denote the upwardmotions) and (b) multilevel westerly wind (u) averaged over 17–22°N. (c and d) Similar to Figures 6a and 6b, respectively,but for the temporal correlation coefficient between the MA mean index of southern Indochina biomass burning activities(defined in Figure 5) and the selected �ω and u averaged over 10–15°N. The time period of correlation is from 2005 to2015. The topography is marked. Only the areas with significant changes (at the 90% confidence interval) are shaded.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,023

(Figure 6b), can then help transport the PM10 from northern Indochina to Taiwan. In comparing Figures 6aand 6b, there is no clear positive relationship between the change in southern Indochina's biomass burningactivities and the change in local upward motion (Figure 6c) or the change in the westerly jet over the areabetween Indochina and Taiwan (Figure 6d). As a result of the differences between Figures 6a and 6b andFigures 6c and 6d, the change in southern Indochina's biomass burning is not as important as the changein northern Indochina's biomass burning in terms of the change of PM10 at LABS in Taiwan.

Notably, in addition to the interannual variation, the PM10 at LABS in Taiwan seems to be modulated by adecreasing trend (see Figure 3). However, the time period of 2007–2015 is too short to reasonably identifythe existence of such a trend. A further examination on this issue is suggested in the future when a longerperiod of data is available.

4. Discussion

From section 3, it can be concluded that the change of the India-Burma Trough plays an important role inmodulating the interannual variation of biomass burning activities over northern Indochina and its impacton PM10 at LABS in Taiwan. Thus, it is important to understand what might cause the change in theIndia-Burma Trough on the interannual time scale. The recent study of Li and Zhou [2016] suggested thatthe interannual variation of the wintertime (December to the following February) India-Burma Trough is partof the South Asian jet wave train and is modulated by El Niño–Southern Oscillation (ENSO) via the PhilippineSea anticyclone. Here we will clarify if the change of the springtime (March and April) India-Burma Trough isalso connected to the change of the South Asian jet wave train and ENSO.

Figure 7 shows the correlation map of the 700 hPa vorticity [i.e., ζ (700 hPa)] and the wind vectors associatedwith the index of northern Indochina's biomass burning activities (defined in section 2). It is interesting tonote that ζ (700 hPa) over the area of the India-Burma Trough is positively correlated with ζ (700 hPa) overthe western Mediterranean Sea and Saudi Arabia and negatively correlated with ζ (700 hPa) over the easternMediterranean Sea and the Arabian Sea. A better illustration of such a wave train can be revealed in the cor-relation map of the 700 hPa stream function [i.e., ψ (700 hPa)] with the northern Indochina's biomass burningindex, as shown in Figure 8a. By comparing this springtime wave train pattern (Figure 8a) with the wintertimewave train pattern identified in Li and Zhou [2016], it is noted that the former also appeared along the SouthAsian jet but with different variation centers. Such a difference might be attributed to the seasonal variationof the South Asian jet.

Previous studies suggested that the propagation of the Rossby wave from the Mediterranean Sea to the Bayof Bengal is one of the key mechanisms for the enhancement of the India-Burma Trough in wintertime [Suo

Figure 7. Correlation maps of the 700 hPa wind (vectors) and vorticity (shading) fields with the MAmean index of northernIndochina biomass burning activities (defined in Figure 4). The time period used for correlation is from 2005 to 2015. Onlythe areas with significant changes (at the 90% confidence interval) are shown.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,024

et al., 2008; Li and Zhou, 2016]. To clarify if this is also true for the enhancement of the India-Burma Trough inthe springtime, we used the area-averaged ζ (700 hPa) over the western Mediterranean Sea (37.5°N–42.5°N,2.5°W–2.5°E) as an index to correlate with ψ (700 hPa) for March and April of 2005–2015. As revealed inFigure 8b, the wave train pattern with an upstream source over the western Mediterranean Sea propagatesdownstream across the Mediterranean Sea, Saudi Arabia, and Arabian Sea to the area covering the northernBay of Bengal, northern Indochina, and south China. This pattern is consistent with that shown in Figure 8a,suggesting that the propagation of such a wave from the western Mediterranean Sea can modulate thechange of the India-Burma Trough in the springtime.

We also examined the possible modulation from the downstream ENSO signal to the India-Burma Trough byusing the wintertime sea surface temperature (SST) averaged over the Niño3.4 region as an index to correlatethe following springtime ψ (700 hPa); the result is given in Figure 8c. It is found that the change of wintertimeSST over the Niño3.4 region is positively related to the change of the following springtime anticyclone overthe area covering most of the South China Sea (hereafter, South China Sea anticyclone (SCSA)). Consistentwith this finding, the bubble chart given in Figure 9 shows that during an El Niño (La Niña) event, a stronger(weaker) SCSA in the following springtime can be expected [e.g., Wang and Zhang, 2002]. As noted from

Figure 8. Correlation maps of the MA (March and April) mean stream function ψ at 700 hPa with (a) the MAmean biomassburning activities averaged over 17–22°N, 97–102°E; (b) the MA mean 700 hPa vorticity averaged over 37.5°W–42.5°E,2.5°W–2.5°E; and (c) the previous winter (DJF, year–1) sea surface temperature averaged over the Niño3.4 region. The timeperiod used for correlation is from 2005 to 2015. Only the areas with significant changes (at the 90% confidence interval)are shaded.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,025

Figures 1d and 7, the existence of SCSA can help strengthen the westerly jet ahead of the India-Burma Trough[e.g.,Wang et al., 2011]. In other words, the change in the springtime India-Burma Trough can be modulatedby the previous winter's ENSO signal from the tropical East Pacific via the SCSA. To better illustrate therelationship between the SCSA and ENSO, a longer time period (1979–2015) of NCEP-DOE Reanalysis 2 datais used for constructing Figure 9. Details of how the SCSA is established during El Niño development can befound in Wang and Zhang [2002].

Finally, we compared the changes of the India-Burma Trough, SCSA, and biomass burning activities overnorthern Indochina by the bubble chart given in Figure 9. It is noted that in the years with strengtheningin both the India-Burma Trough and SCSA (i.e., 2007 and 2010), the amount of biomass burning over northernIndochina is greater than other years with an increase in the India-Burma Trough (i.e., 2012). Consistent withthis finding, in the years with weakening in both the India-Burma Trough and SCSA (i.e., 2008 and 2011), theamount of biomass burning over northern Indochina is lower than in other years with a decrease in the India-Burma Trough (i.e., 2005 and 2015). Based on these results, it is suggested that even though the changes ofthe India-Burma Trough do not always follow the SCSA changes (or the ENSO signal changes), monitoring theSCSA changes (or the ENSO signal changes) does aid in the prediction of the interannual variation of biomassburning over northern Indochina and its impact on PM10 in Taiwan. For instance, weakened SCSA and abnor-mally low biomass burning in 2011 (Figures 2a and 3a) are found consistently with the unusually high 2011premonsoon rainfall, particularly in March, over northern Thailand [Promchote et al., 2015]. This likelysuggests that high precipitation can reduce the burning activities over northern Indochina.

It should be noted that although our study reveals the correlations among biomass burning, the India-BurmaTrough, and SCSA, there are some additional factors, such as land use change [Reid et al., 2013], that might beinvolved in the changes of biomass burning. For instance, maize-cultivated lands were largely expanded in2007 over the mountainous terrain in northern Thailand, Myanmar, and Lao, promoted by the high grainprice [Food and Agriculture Organization of the United Nations, 2016] and private companies; crop residueburning was probably largely increased in this year.

Figure 9. Bubble plot of the springtime (March and April mean) India-Burma Trough index (defined as the area-averaged700 hPa vorticity, ζ , over 20–30°N, 85–110°E) and the corresponding South China Sea anticyclone index (defined as the“negative” 700 hPa vorticity, �ζ , area averaged over 5–15°N, 95–120°E) during 1979–2015. The explanation for examiningthe changes over the longer time period (1979–2015) is given in the manuscript. The orange (blue) bubbles are the positive(negative) sea surface temperature anomalies (SSTAs) over the Niño3.4 region in the previous winter, with the size of thebubbles indicating the magnitude of the SSTA. The red (blue) open circles are the years with maximum (minimum) PC1values of Figure 2.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,026

5. Summary

In this study, the spatial-temporal char-acteristics of the interannual variationof biomass burning in Indochina andits downwind impact on PM10 at LABS(a high-mountain station) in Taiwan areexamined. Diagnoses show that bio-mass burning activities in Indochinavary with two geographical regions: aprimary region in northern Indochinaand a secondary region in southernIndochina. PM10 at LABS in Taiwan ismainly associated with the interannualvariation of biomass burning overnorthern Indochina but not oversouthern Indochina.

Changes in biomass burning overnorthern Indochina and southernIndochina are found to be modulatedby different types of atmospheric circu-lation changes. It is noted that the roleof the India-Burma Trough is importantto the changes of northern Indochina'sbiomass burning activities and its asso-ciated aerosol transport pattern. Assummarized in Figure 10, our resultsshow that in the years with an intensi-

fied India-Burma Trough, an increase in northwesterly wind over the southwest side of the Tibetan Plateau(i.e., behind the India-Burma Trough) can advect dryer air from higher latitude to northern Indochina. Theassociated dryer condition promotes local biomass burning activities. Furthermore, ahead of the intensifiedIndia-Burma Trough, the associated increase in upward motion helps uplift the polluted air to the free tropo-sphere, in which the associated increase in the westerly jet transports the PM10 from northern Indochina toTaiwan. As for the change in the springtime India-Burma Trough, we noted that it is not only teleconnectedwith the upstream wave trains from the western Mediterranean Sea but also modulated by the previous win-ter's ENSO signal from the tropical East Pacific via the South China Sea anticyclone.

The findings of this study are an important step forward in understanding the role of the India-Burma Troughfor a better prediction of the interannual variation of northern Indochina's biomass burning activity and itsdownwind impact on PM10 at LABS in Taiwan. However, for the shorter time scale variations, the role ofthe India-Burma Trough in the modulation of biomass burning in Indochina and its downwind impacts mightbe different than that revealed in this study. Further examinations are suggested in the future to explore thisissue in detail.

ReferencesChen, T. C., M. C. Yen, and M. Murakami (1988), The water vapor transport associated with the 30–50 day oscillation over the Asian monsoon

regions during 1979 summer, Moon Weather Rev., 116, 1983–2002.Cheng, F. Y., Z. M. Yang, C. F. O. Yang, and F. Ngan (2013), A numerical study of the dependence of long-range transport of CO to a mountain

station in Taiwan on synoptic weather patterns during the Southeast Asia biomass burning season, Atmos. Environ., 78, 277–290.Chuang, M. T., et al. (2015), Simulating the transport and chemical evolution of biomass burning pollutants originating from Southeast Asia

during 7-SEAS/2010 Dongsha experiment, Atmos. Environ., 112, 294–305.Dong, X., and J. S. Fu (2015), Understanding interannual variations of biomass burning from Peninsular Southeast Asia, part II: Variability and

different influences in lower and higher atmosphere levels, Atmos. Environ., 115, 9–18.Duncan, B. N., R. V. Martin, A. C. Staudt, R. Yevich, and J. A. Logan (2003), Interannual and seasonal variability of biomass burning emissions

constrained by satellite observations, J. Geophys. Res., 108(D2), 4040, doi:10.1029/2002JD002378.Food and Agriculture Organization of the United Nations (FAO) (2016), FAOSTAT data base. [Available at http://faostat.fao.org].

Figure 10. Schematic diagram for illustrating how an intensified India-Burma Trough modulates the occurrence of biomass burning in north-ern Indochina and its downwind impact on Taiwan. The numbers of 1~3represent the order of the process. Step 1: an increase in northwesterlywind, behind (to the west) the intensified India-Burma Trough, advectsthe dryer air from higher latitude to northern Indochina. Step 2: the dryercondition (i.e., less precipitation and moisture divergence) promotes localbiomass burning activities. Step 3: the increase in upward motion, ahead(to the east) of the intensified India-Burma Trough, helps uplift the pol-luted air to the free troposphere, in which the associated increase in thewesterly jet transports the PM10 from northern Indochina to Taiwan.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,027

AcknowledgmentsThe NCEP-DOE Reanalysis 2 data set wasprovided by the NOAA/OAR/ESRL PSD,Boulder, CO, USA, from their website athttp://www.esrl.noaa.gov/psd/. TheMODIS fire count data and the GFED4sbiomass burning emissions wereprovided by http://reverb.echo.nasa.gov and http://globalfiredata.org/data.html, respectively. The authors thankRitesh Gautam and anonymousreviewers for their comments andsuggestions which greatly improved themanuscript. This research wassupported by the Ministry of Scienceand Technology of Taiwan under MOST104-2111-M-003-001, MOST 105-2119-M-003-002, and MOST 105-2625-M-003-002. S.-H. Wang was supported byMOST 104-2111-M-008-009. N.-H. Linwas supported by MOST 103-2111-M-008-001 and EPA-105-U1L1-02-A046funded by Taiwan EnvironmentalProtection Administration.

Gautam, R., N. C. Hsu, T. F. Eck, B. N. Holben, S. Janjai, T. Jantarach, S. C. Tsay, and W. K. Lau (2013), Characterization of aerosols over theIndochina peninsula from satellite-surface observations during biomass burning pre-monsoon season, Atmos. Environ., 78, 51–59.

Giglio, L. (2013), MODIS collection 5 active fire product user's guide version 2.5. 61 pp. [Available at http://modis-fire.umd.edu/files/MODIS_Fire_Users_Guide_2.5.pdf.]

Giglio, L., J. Descloitres, C. O. Justice, and Y. J. Kaufman (2003), An enhanced contextual fire detection algorithm for MODIS, Remote Sens.Environ., 87, 273–282.

Giglio, L., J. T. Randerson, G. R. van der Werf, P. S. Kasibhatla, G. J. Collatz, D. C. Morton, and R. S. DeFries (2010), Assessing variability andlong-term trends in burned area by merging multiple satellite fire products, Biogeosciences, 7, 1171–1186.

Giglio, L., J. T. Randerson, and G. R. van der Werf (2013), Analysis of daily, monthly, and annual burned area using the fourth-generationglobal fire emissions database (GFED4), J. Geophys. Res. Biogeosci., 118, 317–328, doi:10.1002/jgrg.20042.

Hannachi, A., I. Jolliffe, and D. Stephenson (2007), Empirical orthogonal functions and related techniques in atmospheric science: A review,Int. J. Climatol., 27, 1119–1152.

Hong, Y., K. L. Hsu, S. Sorooshian, and X. Gao (2005), Improved representation of diurnal variability of rainfall retrieved from the TropicalRainfall Measurement Mission Microwave Imager adjusted Precipitation estimation from Remotely Sensed Information Using ArtificialNeural Networks (PERSIANN) system, J. Geophys. Res., 110, D06102, doi:10.1029/2004JD005301.

Huang, K., J. S. Fu, N. C. Hsu, Y. Gao, X. Dong, S. C. Tsay, and Y. F. Lam (2013), Impact assessment of biomass burning on air quality in Southeastand East Asia during BASE-ASIA, Atmos. Environ., 78, 291–302.

Huang, W. R., and J. C. L. Chan (2012), Seasonal variation of diurnal and semidiurnal variation of rainfall over southeast China, Clim. Dyn., 39,1913–1927.

Huang, W. R., and S. Y. Wang (2014), Impact of land-sea breezes at different scales on the diurnal rainfall in Taiwan, Clim. Dyn., 43, 1951–1963.Huffman, G. J., R. F. Adler, D. T. Bolvin, G. Gu, E. J. Nelkin, K. P. Bowman, Y. Hong, E. F. Stocker, and D. B. Wolff (2007), The TRMM multisatellite

precipitation analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales, J. Hydrometeorol., 8(1),38–55.

Kanamitsu, M., W. Ebisuzaki, J. Woollen, S. K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter (2002), NCEP–DOE. AMIP-II Reanalysis (R-2), Bull. Am.Meteorol. Soc., 83, 1631–1643.

Kim Oanh, N. T., and K. Leelasakultum (2011), Analysis of meteorology and emission in haze episode prevalence over mountain-boundedregion for early warming, Sci. Total Environ., 409, 2261–2271.

Lee, C. T., et al. (2011), The enhancement of PM2.5 mass and water-soluble ions of biosmoke transported from Southeast Asia over theMountain Lulin site in Taiwan, Atmos. Environ., 45(32), 5784–5794.

Li, X. Z., and W. Zhou (2016), Modulation of the interannual variation of the India-Burma Trough on the winter moisture supply oversouthwest China, Clim. Dyn., 46, 147–158.

Lin, N. H., et al. (2013), An overview of regional experiments on biomass burning aerosols and related pollutants in Southeast Asia: FromBASE-ASIA and Dongsha experiment to 7-SEAS, Atmos. Environ., 78, 1–19.

Ou Yang, C. F., N. H. Lin, C. C. Lin, S. H. Wang, G. R. Sheu, C. T. Lee, R. C. Schnell, P. M. Lang, T. Kawasato, and J. L. Wang (2014), Characteristics ofatmospheric carbon monoxide at a high-mountain background station in East Asia, Atmos. Environ., 89, 613–622, doi:10.1016/j.atmosenv.2014.02.060.

Promchote, P., S. Y. S. Wang, and P. G. Johnson (2015), The 2011 great flood in Thailand: Climate diagnostics and Implications from climatechange, J. Clim., 29, 367–379, doi:10.1175/JCLI-D-15-0310.1.

Reid, J. S., et al. (2013), Observing and understanding the Southeast Asian aerosol system by remote sensing: An initial review and analysisfor the Seven Southeast Asian Studies (7SEAS) program, Atmos. Environ., 122, 403–468.

Sanap, S. D., and G. Pandithurai (2015), The effect of absorbing aerosols on Indian monsoon circulation and rainfall: A review, Atmos. Res.,164-165, 318–327.

Shi, Y., and Y. Yamaguchi (2014), A high-resolution and multi-year emissions inventory for biomass burning in Southeast Asia during2001-2010, Atmos. Environ., 98, 8–16.

Sonkaew, T., and R. Macatangay (2015), Determining relationships and mechanisms between tropospheric ozone column concentrationsand tropical biomass burning in Thailand and its surrounding regions, Environ. Res. Lett., 10, 065009, doi:10.1088/1748-9326/10/6/.

Suo, M. Q., and Y. H. Ding (2009), The structure and evolution of the wintertime southern branch trough in the subtropical westerlies [inChinese], Chin. J. Atmos. Sci., 33(3), 425–442.

Suo, M. Q., Y. H. Ding, and J. Y. Wang (2008), Relationship between Rossby wave propagation in southern branch of westerlies and theformation of the southern branch trough in wintertime [in Chinese], J. Appl. Meteorol. Sci. 19(6), 731–740.

Wang, B., and Q. Zhang (2002), Pacific-East Asian teleconnection. Part II: How the Philippine Sea anomalous anticyclone is established duringEl Niño development, J. Clim., 15, 3252–3265.

Wang, S. H., et al. (2015), Vertical distribution and columnar optical properties of springtime biomass burning aerosols over northernIndochina during 2014 7-SEAS campaign, Aerosol Air Qual. Res., 15, 2037–2050, doi:10.4209/aaqr.2015.05.0310.

Wang, T. M., S. Yang, Z. P. Wen, R. G. Wu, and P. Zhao (2011), Variations of the winter India-Burma Trough and their links to climate anomaliesover southern and eastern Asia, J. Geophys. Res., 116, D23118, doi:10.1029/2011JD016373.

Yen, M. C., C. M. Peng, T. C. Chen, C. S. Chen, N. H. Lin, R. Y. Tzeng, Y. A. Lee, and C. C. Lin (2013), Climate and weather characteristics inassociation with the active fires in northern Southeast Asia and spring air pollution in Taiwan during 2010 7-SEAS/Dongsha Experiment,Atmos. Environ., 78, 35–50.

Yu, J.-Y., N.-H. Lin, H.-H. Hsieh, and T.-S. Lin (2008), A cluster analysis of the springtime forward trajectories arising from Southeast Asia and theclimate influence [in Chinese with English abstract], Atmos. Sci., 36, 287–301.

Zhou, T., R. Yu, H. Chen, A. Dai, and Y. Pan (2008), Summer precipitation frequency, intensity, and diurnal cycle over China: A comparison ofsatellite data with rain gauge observations, J. Clim., 21, 3997–4010.

Zhou, W., J. C. L. Chan, W. Chen, J. Ling, J. G. Pinto, and Y. P. Shao (2009), Synoptic-scale controls of persistent low temperature and icyweather over southern China in January 2008, Moon Weather Rev., 137, 3978–3991.

Zong, H. F., C. L. Bueh, J. Wei, and L. T. Chen (2012), Intensity of the trough over the Bay of Bengal and its impact on the southern Chinaprecipitation in winter, Atmos. Ocean Sci. Lett., 5(3), 246–251.

Journal of Geophysical Research: Atmospheres 10.1002/2016JD025286

HUANG ET AL. INDOCHINA BIOMASS BURNING AND ITS IMPACT 10,028